Accented Indian english ASR: Some early results

The problem of the effect of accent on the performance of Automatic Speech Recognition (ASR) systems is well known. In this paper, we study the effect of accent variability on the performance of the Indian English ASR task. We evaluate the test vocabularies on HMMs trained on (a) Accent specific training data (b) Accent pooled training data which combines all the accent specific training data (c) Accent pooled training data of reduced size matching the size of the accent specific training data. We demonstrate that the accent pooled training set performs the best on phonetically rich isolated word recognition task. But the accent specific HMMs perform better than the reduced accent pooled HMMs, indicating a possible approach of using a first stage accent identification to choose the correct accent trained HMMs for further recognition.

[1]  Josef G. Bauer.  On the choice of classes in MCE based discriminative HMM-training for speech recognizers used in the telephone environment , 2001, INTERSPEECH.

[2]  Philip C. Woodland,et al.  The use of accent-specific pronunciation dictionaries in acoustic model training , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[3]  Josef G. Bauer,et al.  Triphone tying techniques combining a-priori rules and data driven methods , 2001, INTERSPEECH.

[4]  John H. L. Hansen,et al.  Language accent classification in American English , 1996, Speech Commun..

[5]  Chao Huang,et al.  Automatic accent identification using Gaussian mixture models , 2001, IEEE Workshop on Automatic Speech Recognition and Understanding, 2001. ASRU '01..